Structurally self-anchoring bipolar membranes with interlocked interfaces for coupled hydrogen production and biomass electrosynthesis

Abstract

The development of efficient and multifunctional electrochemical systems is critical for advancing sustainable hydrogen production and green chemical synthesis. Herein, we report a structurally self-anchoring bipolar membrane (SA-BPM) fabricated via a scalable templated casting strategy that enables spontaneous physical interlocking between the cation exchange layer (CEL) and the anion exchange layer (AEL). This architecture significantly enhances interfacial adhesion, mechanical stability, and water dissociation kinetics. When integrated into an electrolyzer, the SA-BPMWE demonstrates a remarkably high current density of 900 mA cm⁻² at 2.92 V, which is more than twice the current density of the S-BMPWE without a self-anchored morphology (S-BPMWE, 374 mA cm⁻²) under the same conditions. Moreover, coupling cathodic hydrogen evolution with anodic 5-hydroxymethylfurfural (HMF) oxidation via the SA-BPM yields a current density of 10 mA cm⁻² at only 0.72 V, enabling simultaneous H₂ generation and production of 2,5-furandicarboxylic acid (FDCA) with 83.96% HMF conversion and 41.34% yield. This work highlights the critical role of interfacial structure in bipolar membranes and establishes a versatile membrane–electrode platform for integrated energy and chemical manufacturing applications.

---

1 Introduction

With the accelerating global transition toward a low-carbon economy, hydrogen (H₂) has emerged as a pivotal energy carrier due to its high energy density and zero-emission characteristics.^1,2 Among various hydrogen production methods, electrochemical water splitting is regarded as one of the cleanest and most scalable approaches when powered by renewable electricity.^3,4 However, its practical deployment is constrained by substantial energy input, largely due to the sluggish kinetics and high overpotential of the oxygen evolution reaction (OER) at the anode.^5–8 Moreover, the generation of low-value oxygen as a by-product limits the economic competitiveness of the overall process.

Bipolar membranes (BPMs), composed of a cation exchange layer (CEL) and an anion exchange layer (AEL), have recently garnered significant interest in electrochemical energy conversion systems due to their unique ability to establish a steep pH gradient across the interface.^9,10 This property enables spatial decoupling of acidic and alkaline reactions, facilitates in situ water dissociation (WD) into H⁺ and OH⁻ ions, and broadens the scope for reaction environment design.^11,12 While BPMs have been extensively used in electrodialysis and CO₂ electroreduction, they are now emerging as a promising platform for water electrolysis, particularly under asymmetric pH conditions that favor efficient ion transport and gas separation.^13–16

A growing strategy to improve the energy efficiency of water electrolysis involves replacing the energy-intensive OER with value-added anodic oxidation reactions.^17,18 In particular, the electrooxidation of biomass-derived 5-hydroxymethylfurfural (5-HMF) into 2,5-furandicarboxylic acid (FDCA) stands out due to its economic advantages and lower thermodynamic potential.^9,19 Coupling anodic 5-HMF oxidation at the anode with the hydrogen evolution reaction (HER) at the cathode not only reduces cell voltage but also enables the co-production of green hydrogen and industrially relevant chemical feedstocks.²⁰

Despite the promise of BPM-coupled electrosynthesis, major challenges persist. Conventional BPMs often suffer from poor interfacial compatibility between the CEL and AEL, leading to high interfacial resistance,²¹ inefficient water dissociation, and membrane delamination during prolonged operation.^22–24 Addressing these limitations requires next-generation BPMs featuring robust interfacial adhesion, optimized ion transport pathways, and stable physicochemical interfaces under operating conditions.^25–28 Recent advances include the use of electro-spun junctions with nanoparticle catalysts,²⁹ soft-lithography patterned microscale arrays,³⁰ and conductive interlayers that enhance local electric field strength to lower WD overpotential.³¹ While these approaches have demonstrated promising performance gains, they often involve complex and cost-intensive fabrication processes, limiting their scalability and practical deployment.³²

In this work, we report a structurally self-anchoring bipolar membrane (SA-BPM) fabricated via a facile, scalable interfacial transfer strategy that induces spontaneous microstructure anchoring at the CEL and AEL junction (Fig. 1). This self-anchoring architecture ensures a large interfacial contact area, promotes water dissociation kinetics, and prevents delamination at high current densities. Integrated into a hybrid electrolysis system coupling hydrogen production with 5-HMF oxidation, the SA-BPM enables significantly lower onset voltage, higher current densities, and superior stability compared to commercial BPMs. Furthermore, the system delivers high FDCA selectivity and prolonged operation at elevated temperatures. These findings not only validate the critical role of physical interfacial structuring in membrane performance but also offer a rational design blueprint for multifunctional BPMs in next-generation electrochemical energy and chemical production systems.

Fig. 1 Schematic illustration of the fabrication process and functional integration of the self-anchoring bipolar membrane (SA-BPM). A patterned AEL is prepared via phase-transfer templating, followed by catalyst loading and CEL spray-coating to form an interlocked interface, enabling efficient water dissociation and coupling of hydrogen evolution with HMF oxidation to FDCA.

2 Experimental section

2.1 Materials

Trifluoromethane sulfonic acid (98% TFSA), trifluoroacetic acid (98% TFA), biphenyl (99.5%), N-methyl-4-piperidone (97% MP), dimethyl sulfoxide (99% DMSO), potassium hydroxide (95% KOH), sodium hydroxide (97% NaOH), iron(II) sulfate heptahydrate (99% FeSO₄·7H₂O), dopamine hydrochloride (98% DA), and copper sulfate pentahydrate (99% CuSO₄·5H₂O) were purchased from Shanghai Macklin Biochemical Co., Ltd. Nickel(II) nitrate hexahydrate (Ni(NO₃)₂·6H₂O, AR) and sulfuric acid (H₂SO₄, AR) were purchased from Sinopharm Chemical Reagent Co., Ltd. Perfluorosulfonic acid ion exchange resin (PFSA) was purchased from Fuzhou Topda New Material Co., Ltd. Bipolar membranes (Fumasep BPM, Fumatech Co., Germany) were used for analysis of water dissociation. Tris (hydroxymethyl) aminomethane buffer solution (10 mM, Tris–HCl, pH 8.5) was provided by Phygene Biotechnology Co., Ltd. Ethanol (100% EtOH) was purchased from Xilong Scientific Co., Ltd. 2,5-Furandicarboxylicacid (98% FDCA), 5- hydroxymethylfurfural (95% HMF), and dimethyl sulfoxide (99% DMSO) were provided by Shanghai Aladdin Biochemical Technology Co., Ltd. Nickel foam (Sci Materials Hub) was used to prepare the HMFOR catalytic electrode. Deionized (DI) water was used throughout the research work.

2.2 Fabrication of bipolar membranes (BPMs)

2.2.1 Synthesis of poly(arylenepiperidinium) (PAP). Poly(arylenepiperidinium) (PAP) was synthesized via a strong acid polymerization reaction between N-methylpiperidinium ketone and aromatic hydrocarbons.^25,33–35 Initially, 6.16 g (0.02 mol) of biphenyl was dissolved in 30 mL of dichloromethane in a 50 mL three-neck round-bottom flask equipped with a magnetic stirrer and a nitrogen inlet. The mixture was stirred at room temperature until a clear and homogeneous solution was formed. The solution was subsequently cooled to 0 °C, and nitrogen gas was purged for 30 minutes to eliminate dissolved oxygen. A mixed acid comprising 3.6 mL of trifluoroacetic acid (TFA) and 36.6 mL of trifluoromethanesulfonic acid (TFSA) was then added dropwise via a dropping funnel. During the polymerization, the solution colour gradually changed from yellow to reddish-purple, and the viscosity increased progressively. The reaction was considered complete when the viscosity became too high for the stir bar to rotate. The resulting viscous mixture was slowly poured into 1 L of 1 M KOH solution under vigorous stirring and allowed to neutralize for 24 hours. The pale-yellow precipitate was collected by filtration, washed repeatedly with deionized (DI) water until the filtrate reached neutral pH, and dried in an oven at 80 °C for 24 hours.

2.2.2 Preparation of the anion exchange layer. The fabrication of the self-anchoring bipolar membrane (SA-BPM) was carried out in three main steps: (1) preparation of the self-anchoring anion exchange layer (SA-AEL), (2) deposition of the catalyst layer (CL), and (3) application of the cation exchange layer (CEL).

To prepare the SA-AEL (anion exchange layer with a self-anchoring morphology), 0.2 g of PAP was combined with 100 µL of 1-bromobutane and 10 mL of dimethyl sulfoxide (DMSO) in a beaker and stirred at 40 °C for 24 hours. The resulting viscous solution was then cast onto an interfacial phase-transfer template (for comparison a smooth AEL, S-AEL, was cast using standard culture dishes) and dried at 80 °C for 24 hours. After drying, the membrane was carefully peeled from the template and immersed in 70 mL of 2 M potassium hydroxide (KOH) solution for ion exchange. Finally, the membrane was rinsed thoroughly with deionized (DI) water to remove residual salts and stored in DI water until further use.

A straightforward method was employed to prepare the catalyst layer (CL) for the bipolar membranes. The catalyst layer was fabricated within a single-sided membrane reactor (Fig. S13). The catalyst layer was fabricated within a single-sided membrane reactor, in which the SA-AEL was positioned with its self-anchoring side exposed to air. A solution containing 30 mg of dopamine (DA) dissolved in 30 mL of pH 8.5 tris–HCl buffer was added to the reactor and left undisturbed for 24 hours to allow the self-polymerization of dopamine. Subsequently, a 5 g L⁻¹ solution of iron(II) sulfate heptahydrate (FeSO₄·7H₂O), prepared in a solvent mixture of ethanol and deionized (DI) water (1 : 3), was introduced into the reactor.^35,36 The membrane remained in the solution for 48 hours, after which it was removed and thoroughly rinsed with DI water.

The self-anchoring bipolar membrane (SA-BPM) was fabricated via a spray-coating technique to ensure uniform deposition and strong interfacial bonding between the layers.³¹ The previously prepared SA-AEL, with the catalyst layer (CL) facing upward, was placed on a hot plate maintained at 55 °C. A 10 g L⁻¹ solution of perfluorosulfonic acid (PFSA) ionomer was then uniformly sprayed onto the exposed catalyst surface using an airbrush. The spraying continued until the cation exchange layer (CEL) reached the target thickness. After deposition, the assembled membrane was thermally treated at 55 °C for a specified period to promote interfacial adhesion and complete curing of the CEL. The resulting membrane exhibited a robust, interlocked structure with excellent mechanical integrity and ion-conductive functionality, suitable for bipolar electrochemical applications.

2.3 Characterization

The chemical structure of the synthesized PAP membrane material was analysed using proton nuclear magnetic resonance (¹H NMR) spectroscopy with DMSO-d₆ as the solvent. ¹H NMR spectra were obtained on an NMR 700 MHz spectrometer (BRUKER AVANCE III, Switzerland). The cross-sectional and surface morphology of the bipolar membrane (BPM) was characterized using field-emission scanning electron microscopy (FE-SEM, Zeiss Gemini 300, Britain), while the elemental composition of the cross-section was further analysed using energy-dispersive X-ray spectroscopy (EDS, Oxford Ultim MAX system, Britain). Before SEM and EDS analysis, the membrane samples were sputter-coated with gold to improve electrical conductivity.

Fourier-transform infrared (FTIR) spectroscopy was performed to identify the main functional groups of PAP using a TENSOR27 spectrometer (Bruker Hong Kong Ltd, Germany). The spectra were recorded in the wavenumber range of 600–4000 cm⁻¹ with potassium bromide (KBr) as the background medium, and the samples were prepared using the pellet technique. The surface morphology of the membrane was analysed using an atomic force microscope (AFM, Dimension Icon, Bruker) in tapping mode.

X-ray photoelectron spectroscopy (XPS) was performed on a Thermo Fisher Scientific Nexsa G2 spectrometer using a monochromatic Al Kα source. XPS analyses were conducted on PAP, PDA–PAP, and FeOOH@PDA–PAP samples. High-resolution spectra were fitted using Thermo Scientific Avantage software.

The ion exchange capacity (IEC) of the anion exchange layer (AEL) was determined using a back-titration method. First, the membrane sample, which was converted into the OH⁻ form, was dried in an oven until equilibrium, and the dry weight was recorded as W_dry. The membrane was then submerged in 30 mL of 0.01 M HCl solution for 24 hours to complete ion exchange. Afterward, a pH meter (Waltham, MA, USA) was used to back-titrate the solution with 0.01 M NaOH. The initial and final volumes of the NaOH solution were recorded, and the IEC of the AEL (IEC_AEL) was calculated using the following equation:

A similar procedure was employed to evaluate the ion exchange capacity of the cation exchange layer (CEL). The CEL was thoroughly dried in an oven for 24 hours and treated in 0.01 M HCl. The exchanged membrane was then titrated using 0.01 M NaOH solution. The IEC of the CEL (IEC_CEL) was calculated using the following equation:

The in-plane conductivity of the OH⁻ form anion exchange layer (AEL) and H⁺ form cation exchange layer (CEL) was measured using a Corrtest (CS310MA) electrochemical workstation. Both AEL and CEL samples were cut into dimensions of 1 cm × 4 cm and placed in a four-electrode cell with platinum (Pt) electrodes for alternating current (AC) impedance testing. The cell containing the membrane samples was immersed in deionized water, and impedance measurements were conducted in galvanostatic mode over frequencies ranging from 1 Hz to 10⁵ Hz to obtain the in-plane ohmic resistance (R). The ionic conductivity (σ) was calculated using the following equation:

where L, T and W represent the length, thickness, and width of the membrane sample, respectively.

The H⁺ form cation exchange layer (CEL) and OH⁻ form anion exchange layer (AEL) were cut into dimensions of 1 cm × 3 cm. The samples were dried to a constant weight, and the dry weight and length were recorded as W_dry and L_dry, respectively. These membrane samples were then immersed in deionized water at temperatures of 30 °C, 50 °C and 70 °C for 24 hours. Afterward, the samples were removed, and excess water was blotted from the surface using filter paper. The wet weight and length were measured and recorded as W_wet and L_wet. The water uptake (WU) and swelling ratio (SP) for both the AEL and CEL were calculated using the following equations:

The bipolar membrane was installed in a four-electrode cell with a Luggin capillary design. The four-electrode H-cell (with two reference electrodes positioned close to each membrane side) was used to quantify the membrane-related potential drop, thereby minimizing the influence of electrode polarization and reducing solution iR contributions when probing intrinsic water-dissociation behavior. Two platinum plate electrodes were used as the working and counter electrodes, while two Ag/AgCl reference electrodes were placed in the Luggin capillary tube and connected to the sensor. Linear sweep voltammetry (LSV) measurements were performed using a Corrtest electrochemical workstation (CS310MA) with dynamic current scanning at a rate of 2 mA s⁻¹. All experiments were conducted at room temperature in a 1 M Na₂SO₄ solution. The experiment was terminated once the scan reached 3 V or when the upper voltage limit of the electrochemical workstation was reached. Electrochemical impedance spectroscopy (EIS) was performed using the same setup, with data collected in constant current mode at frequencies ranging from 1 Hz to 10⁵ Hz in galvanostatic (20 mA cm⁻²) mode. The resulting Nyquist plots were fitted using ZView software based on equivalent circuits.

A two-electrode zero-gap flow electrolyzer was employed to assess the device-level performance of SA-BPMs under application-relevant conditions, including high-current operation and durability, using the same active membrane area as in the H-cell and reporting current densities normalized to mA cm⁻². The electrolyzer comprises electrodes, a bipolar membrane, bipolar plates with flow fields, and gaskets. During assembly, the cation exchange layer (CEL) is positioned facing the cathode, and the anion exchange layer (AEL) faces the anode. The temperature of the cell is controlled by a tubular heater and thermocouples. Inside the cell, 0.5 M H₂SO₄ is introduced into the cathode chamber, while 1 M KOH is injected into the anode chamber. Continuous circulation of the solutions in both chambers is facilitated by a peristaltic pump. A linear sweeping voltammetry (LSV) curve is obtained by applying a scan rate of 10 mV s⁻¹ over a voltage range of 0 to 3 V at a controlled temperature.

The electrochemical setup for the HMFOR coupling test was molded. For this evaluation, a highly efficient and easily prepared NiCuO_x/NF electrode was selected as the anode catalyst,³⁷ while the SA-BPM served as the separator. The anodic electrolyte consisted of 30 mL of 1 M KOH containing 10 mM HMF, while the cathodic electrolyte comprised 30 mL of 0.5 M H₂SO₄. Linear sweep voltammetry (LSV) measurements were conducted at a scan rate of 5 mV s⁻¹. The products of HMF electrooxidation were analysed by high-performance liquid chromatography (HPLC). Reaction samples were collected at intervals corresponding to every 25 C of charge passed, with 10 µL of the reaction solution diluted in 490 µL of ultrapure water.³⁸ The resulting solutions were filtered using aqueous-phase filter heads to obtain the testing samples. Chromatographic analyses were performed with a Shimpack GWS 5 µm C18 column (4.6 mm × 150 mm), with a 10-minute separation time per analysis. The detection wavelength was set to 265 nm, and the mobile phase consisted of methanol and a 5 mM ammonium formate solution (methanol-to-ammonium formate volume ratio: 3 : 7). The total flow rate of the mobile phase was maintained at 0.6 mL min⁻¹. Standard curves were generated using pure reactant and target product solutions (HMF and FDCA) for product identification and quantitative analysis. The conversion rate of HMF, the yield of FDCA, and the faradaic efficiency were calculated using the following equations:

where n is the molar concentration of relative chemicals, F is the Faraday constant (96 485 C mol⁻¹), and Q is the electrolysis charge, C. Repeatability was evaluated by independent experiments performed on separately assembled cells/membranes under identical conditions.

3 Results and discussion

3.1 Interfacial structure and morphological characterization

In this approach, poly(arylethylidene) (PAP) ^25,33,39 and perfluorosulfonic acid (PFSA) were selected as the anion exchange layer (AEL) and cation exchange layer (CEL) respectively, owing to their excellent conductivity and chemical stability (Fig. S3 and Table S1). The molecular structure of the PAP precursor was confirmed by ¹H NMR and FTIR spectroscopy (Fig. S1 and S2).⁴⁰ Surface observations revealed that the AEL featuring a templated microstructure exhibited a matte, low-gloss appearance, in contrast to the reflective and smooth morphology of the unpatterned AEL (Fig. S4), indicative of surface topology differences. SEM images confirm that the templated AEL possesses pronounced micro- and nano-scale undulations, while the unpatterned AEL remains flat and featureless (Fig. S5). AFM analysis further reveals that root-mean-square roughness (R_a) of the patterned AEL reaches 60.6 nm, more than 600-fold greater than that of the smooth AEL with R_a = 0.10 nm (Fig. S6). This substantial increase in surface roughness, achieved via templated casting, is expected to enlarge the interfacial contact area and enhance local electric field intensity—both of which are critical for promoting water dissociation.^41,42 XPS survey spectra show the emergence of Fe signals only after FeOOH deposition, and the Fe 2p core-level spectrum further displays Fe(III) oxyhydroxide-like features, supporting the successful construction of the FeOOH-based catalyst interlayer (Fig. S14 and S15).

To elucidate the effect of interfacial structure on bipolar membrane performance, we systematically compared the surface and cross-sectional features of the self-anchored bipolar membrane (SA-BPM) and the conventional smooth-interface BPM (S-BPM), as shown in Fig. 2. Cross-sectional SEM imaging (Fig. 2a) reveals a tightly integrated layer in the SA-BPM, characterized by interlocked features between the CEL and AEL, whereas the S-BPM exhibited a flat and mechanically vulnerable interface (Fig. S5). Elemental mapping of nitrogen and fluorine (Fig. 2b and c) effectively delineated the spatial distribution of the AEL and CEL, confirming a structurally coherent junction and successful transfer of the self-anchored morphology. Together, these results demonstrate that the self-anchored architecture not only enhances mechanical robustness by preventing delamination but also provides a structurally optimized environment for interfacial catalysis, laying the groundwork for improved water dissociation kinetics and operational stability in BPM-based electrolysis.

Fig. 2 Structural and electrochemical characterization of self-anchoring and smooth-interface bipolar membranes. (a) Cross-sectional SEM image of the SA-BPM. (b and c) Elemental mapping of fluorine and nitrogen indicating the CEL and AEL regions, respectively. (d) Schematic comparison of the water dissociation layer in the SA-BPM and S-BPM. (e and f) FE-SEM images of the patterned AEL with CL and the smooth AEL with CL, respectively. (g) I–V polarization curves of the SA-BPM and S-BPM; the inset shows the transmembrane voltage at 100 mA cm⁻². (h) Nyquist plots and corresponding electrical equivalent circuit fitting curves of the SA-BPM and S-BPM. The inset shows the equivalent circuit model, which is composed of a solution resistance (R_Ω), a water dissociation resistance (R_wd) and a constant phase element (CPE) connected in parallel, along with a Gerischer impedance (G). In the equivalent circuit, R_wd denotes the resistance associated with the water-dissociation reaction at the bipolar junction (often reported as R_ct in general EIS nomenclature) and is used here to emphasize the junction-specific process.

3.2 Water dissociation performance

To elucidate the role of interfacial architecture in water dissociation efficiency, we compared the catalyst loading behaviour between the templated (patterned) and the conventional (smooth) AEL. As illustrated in Fig. 2d, the SA-BPM features an interlocked interface composed of the CEL, catalyst layer (CL), and a textured AEL, featuring microstructural undulations introduced via templating. In our design, a thin polydopamine (PDA) was first applied to enhance catalyst nucleation, followed by in situ growth of FeOOH as the catalytic layer for water dissociation. This interfacial texture not only increases the contact area but also facilitates anchoring of catalytic nanoparticles.

SEM images in Fig. 2e and f reveal that the patterned AEL facilitates uniform, conformal dispersion of the FeOOH catalyst across the undulating surface. In contrast, the smooth AEL lacks topography texture for anchoring—displays sparse and uneven catalyst coverage, with visible regions of catalyst aggregation and exposed membrane surface. This observation highlights the importance of microstructure interfaces in achieving homogeneous catalyst immobilization, which is critical for stable and efficient interfacial catalysis.^21,26,43

The impact of interfacial morphology on water dissociation behaviour was further assessed by linear sweep voltammetry (LSV) in a 4-electrode cell setup (SI Fig. S7). As shown in Fig. 2g, the SA-BPM exhibited a substantially lower transmembrane voltage of 1.38 V at 100 mA cm⁻² (U₁₀₀),^44,45 compared to 1.93 V for the S-BPM, despite identical membrane materials and catalyst loadings. This 550 mV voltage reduction underscores the dominant contribution of interfacial morphology to electrochemical performance. Quantitatively, AFM topography (Fig. S6) shows that the patterned AEL used for SA-BPM fabrication exhibits a markedly higher roughness (R_a = 60.6 nm) than the smooth AEL (R_a = 0.10 nm), supporting an enlarged effective interfacial contact area and increased accessibility of interfacial sites at the bipolar junction. Moreover, the onset voltage for water splitting with the SA-BPM was only 0.79 V, outperforming the S-BPM (1.11 V) under identical conditions (SI Fig. S8). Here, the onset voltage was defined as the applied voltage at which the interfacial water dissociation reaction begins to dominate the current generation, and we obtained the onset potential by the intersection points of the tangents seen on the I–V curve; the elevated current observed in the initial low-voltage region is treated as a background contribution (likely from parasitic ionic conduction/leakage in the H-cell configuration) and is therefore not used for onset determination. The onset potential of the SA-BPM is 320 mV lower than that of the S-BPM, indicating an effective increase in the interlayer contact area by the self-anchored interface. The increased interfacial roughness and physical interlocking in the SA-BPM promotes intensified local electric fields, improved ion transport, and more efficient water dissociation kinetics.

To further investigate the underlying mechanism, electrochemical impedance spectroscopy (EIS) measurements were conducted (Fig. 2h). The Nyquist plots show that the SA-BPM exhibits a smaller semicircle diameter, corresponding to a reduced interfacial resistance. Equivalent circuit modelling revealed a water dissociation resistance (R_wd) of only 4.8 Ω for the SA-BPM, in contrast to 35 Ω for the S-BPM. The pronounced reduction in R_wd highlights the role of the self-anchored microstructure in facilitating faster ion dissociation and charge transfer across the CEL–AEL interface. Mechanistically, the textured interface increases the effective electrochemical area and catalytic site density, while also promoting improved separation of protons and hydroxide ions.

Beyond structural integration, we suppose that the self-anchored interface provides dual functional advantages. First, it significantly enhances mechanical adhesion at the bipolar junction, effectively mitigating delamination and blistering under long-term electrochemical operation. Second, the patterned interface offers abundant active sites and intensified local electric fields, thereby promoting water dissociation kinetics. This dual role of mechanical and functional optimization positions the SA-BPM as a promising platform for next-generation BPM applications, particularly in scenarios requiring high current density operation and membrane-electrode compatibility. Looking forward, integrating more active water-dissociation catalysts with scalable, automated fabrication methods (e.g., ultrasonic spray-coating) should enable large-area, uniform SA-BPM production while further boosting junction kinetics.

These results clearly demonstrate that rational interface engineering particularly through structural anchoring is an effective and scalable strategy to enhance the kinetics and operational stability of bipolar membranes.

3.3 Coupled water electrolysis performance

To evaluate the practical applicability of the self-anchored bipolar membrane (SA-BPM), a series of polarization and stability tests were conducted using a full-cell electrolyzer configuration, benchmarked against both a smooth BPM (S-BPM) and a commercial Fumasep FBM®. For device benchmarking in the zero-gap electrolyzer, the Fumasep FBM was used as a commercial reference, while the S-BPM was used as a structure-matched laboratory control in the H-cell measurements. As shown in Fig. 3a, the electrolyzer was assembled with commercial non-precious Ni–Mo-based electrodes and asymmetric electrolytes (0.5 M H₂SO₄ ‖ 1 M KOH). Polarization curves recorded at 30 °C, as shown in Fig. 3b, clearly demonstrate that the SA-BPM delivers significantly improved current density across the entire voltage range compared to the S-BPM and FBM. This kinetic advantage is further reflected in high-rate operation, where the SA-BPM achieves 200 mA cm⁻² at just 1.73 V, while the FBM requires 2.94 V to reach the same current density. Fig. 3c presents the dynamic current response of the SA-BPM and commercial Fumasep FBM across a current density range of 0–1000 mA cm⁻² at 50 °C. The SA-BPM system achieved a current density of 1000 mA cm⁻² at a cell voltage of 3.82 V. In contrast, due to limitations of the electrochemical workstation, the voltage required for the FBM to reach 1000 mA cm⁻² could not be determined. Notably, even at a high applied voltage of 9.8 V, the FBM system only reached 638 mA cm⁻², underscoring the substantial performance gap between the two membranes. To decouple structural effects from catalytic activity, we further tested a catalyst-free SA-BPM (without the FeOOH/PDA@FeOOH interlayer) under the same electrolyzer conditions, confirming that the interlocked junction alone provides a measurable contribution to the enhanced water dissociation performance (Fig. S16). This enhanced electrolysis performance is attributed to the increased interfacial area and reduced water dissociation resistance provided by the self-anchoring junction architecture.

Fig. 3 (a) Schematic of a bipolar membrane water electrolyzer in aqueous solutions with 1 M KOH as the anolyte and 0.5 M H₂SO₄ as the catholyte; the anode and cathode catalysts are commercial Ni/Mo-based catalysts for the OER and HER, respectively. (b) I–V curves of the SA-BPM, S-BPM and Fumasep FBM at 30 °C. The inset shows the cell voltage at 200 mA cm⁻². (c) Comparison of I–V curves of the SA-BPM and Fumasep FBM up to1000 mA cm⁻² at 50 °C. (d and e) I–V curves of the SA-BPM and Fumasep FBM at 50 °C and 70 °C, and the corresponding cell voltage at 200 mA cm⁻². (f) Long-term water electrolysis stability of the SA-BPM and S-BPM (at a current density of 100 mA cm⁻²).

To further explore temperature effects on electrochemical performance, current–voltage (I–V) characteristics were measured at elevated temperatures (Fig. 3d). The SA-BPM exhibited excellent thermal responsiveness: at 70 °C, it reaches 900 mA cm⁻² at only 2.92 V, compared to 333 mA cm⁻² for the FBM under identical conditions and around 400 mA cm⁻² for the S-BPM (Fig. S9). This improvement arises from both improved ion mobility and facilitated water dissociation kinetics at higher temperatures. Furthermore, Fig. 3e compares the cell voltages required to reach 200 mA cm⁻² under different conditions. The SA-BPM showed a significant voltage reduction from 1.34 V (50 °C) to 1.15 V (70 °C), whereas the reduction for the FBM was more modest. These results underscore the strong compatibility of the SA-BPM with high-current and high-temperature operation, satisfying key performance metrics for practical hydrogen production systems.

Beyond short-term metrics, long-term durability is crucial for membrane viability in electrochemical systems. Fig. 3f presents the chronoamperometric stability test at 100 mA cm⁻² over 110 h. The S-BPM system exhibited significant voltage drift (0.8 V), indicating poor interfacial stability, due to delamination and swelling mismatch. In contrast, the SA-BPM maintained a stable voltage profile throughout the test with only a minor increase of 0.26 V. Cross-sectional SEM images of the membrane after operation (inset) illustrate the difference in mechanical behaviour: the smooth S-BPM interface is susceptible to mechanical failure, while the physically interlocked interface in the SA-BPM preserves integrity under hydration and electrochemical stress. Overall, the long-term stability of BPMWEs can be influenced by electrode/catalyst-layer durability (e.g., corrosion or leaching), chemical stability of the AEL/CEL and junction/interlayer under sustained pH gradients, and mechanical water-management effects (swelling/compression and gas accumulation) that alter interfacial contact resistance.

3.4 Coupling bipolar membrane water electrolysis with the HMFOR

To demonstrate the extended functionality of the self-anchored bipolar membrane (SA-BPM) beyond traditional water electrolysis, we constructed a hybrid electrolyzer coupling cathodic hydrogen evolution (HER) with anodic oxidation of biomass-derived 5-hydroxymethylfurfural (HMF), as shown in Fig. 4a. In this system, a Ni–Mo-based non-precious catalyst was employed at the cathode, while a NiCuO_x-modified nickel foam served as the HMFOR catalyst.^37,46 The SA-BPM serves as a highly stable and catalytically active ion-conducting separator. During operation, protons generated at the bipolar interface are consumed at the cathode for hydrogen generation, while HMF is oxidized to 2,5-furandicarboxylic acid (FDCA) at the anode, forming a coupled energy-chemical production platform.^17,47–49 The use of a self-anchoring bipolar membrane is particularly advantageous in such systems due to its ability to maintain tight interfacial contact and resist structural degradation under high current conditions and prolonged operation.

Fig. 4 (a) Schematic illustration of the HER coupled with the HMFOR. (b) Comparison of I–V curves equipped with the SA-BPM and Fumasep FBM with and without 10 mM HMF. (c) I–t curve for the HMFOR coupled with the HER. (d) HMF conversion efficiency. (e) The HMF conversion, FDCA yield, and FE formation in the electrolyzer.

To evaluate system performance, linear sweep voltammetry (LSV) was performed under identical conditions for HER//HMFOR and HER//OER operation. As shown in Fig. 4b, the HER//HMFOR system based on the SA-BPM achieved a current density of 10 mA cm⁻² at an exceptionally low cell voltage of only 0.72 V. This represents a 190 mV reduction in voltage compared to the HER//OER system using the same electrode materials. In comparison, employing commercial Fumasep FBM® under identical experimental conditions required a significantly higher voltage of 1.63 V to achieve the same current density, nearly twice the voltage required by the SA-BPM system. Additionally, the LSV curve further reveals a characteristic peak around 1.42 V, corresponding to the voltage region where the competing OER surpasses the HMFOR in current contribution.^17,38 This dramatic improvement highlights the intrinsic advantages of the SA-BPM interface, specifically its superior water dissociation kinetics and efficient ionic transport properties. To quantitatively evaluate the coupling efficiency of the SA-BPM-based system, electrocatalytic HMFOR experiments were performed at a constant applied potential of 1.4 V, with product evolution monitored by high-performance liquid chromatography (HPLC, Fig. S10). As shown in Fig. 4c, the concentration of HMF steadily declined over the electrolysis duration, accompanied by a gradual decrease in current density due to reactant consumption.

Product distribution and selectivity were analysed to quantify the system's electro-synthetic efficiency. As shown in Fig. 4d, HMF conversion efficiency steadily increases with charge input, reaching 83.96% at 175 C. Correspondingly, Fig. 4e shows that the yield of FDCA increased with increasing charge, reaching 41.34%, whereas the faradaic efficiency (FE) decreases at higher charge input, suggesting that an increasing fraction of charge is consumed by competing processes (e.g., side reactions and/or the OER) and that not all converted HMF is directed to FDCA. After conducting three repeated tests under the same conditions, the relative differences (compared with the mean values) in FDCA yield, HMF conversion and faradaic efficiency (FE) were 0.48%, 0.94% and 0.98%, demonstrating the reproducibility of the HMFOR coupled system (Fig. S17) The increasing trend of HMF conversion and FDCA yield confirms the effective and selective oxidation catalysed by NiCuO_x. In contrast, tests using Ni(OH)₂ (ref. 50) as a benchmark catalyst resulted in an HMF conversion rate of 56.69% and an FDCA yield of only 10.01% (Fig. S11 and S12), further highlighting the synergistic role of catalyst selection and membrane architecture.

Although this work focuses on the role of the self-anchoring bipolar junction in enabling coupled HER|HMFOR operation, the HMFOR pathway has been extensively studied on Ni-based catalysts in alkaline media. Prior time-resolved product analyses on Ni–Cu catalysts support a sequential oxidation route HMF → HMFCA → FFCA → FDCA, where HMFCA and FFCA are the major detectable intermediates, while the DFF route is typically disfavored under alkaline electro-oxidation conditions. In our system, only HMF conversion and FDCA yield were quantified; therefore, the conversion–yield gap likely reflects the presence of partially oxidized intermediates and/or alkaline degradation/condensation products.⁵¹

Notably, the electrolyte turned yellow after the coupling reaction, which is consistent with HMF degradation/condensation in strongly alkaline media and the formation of humins-like by-products. To enhance FDCA selectivity, practical strategies include employing more selective HMFOR catalysts (especially to accelerate the late-stage oxidation toward FDCA while suppressing the OER), optimizing the operating cut-off charge/current regime to avoid substrate-depleted conditions, and improving flow-cell operation (e.g., higher flow rate or single-pass/continuous feeding) to reduce residence time and mitigate chemical degradation. These results underscore the capability of the SA-BPM to serve not only as a mechanically robust separator but also as a functional interface in coupled electrolysis–electrosynthesis systems, enabling concurrent hydrogen and high-value chemical production. Long-term coupled durability will be evaluated in future work using continuous-feeding/single-pass flow operation to maintain a quasi-steady HMF concentration, thereby enabling meaningful multi-cycle or extended-time stability assessment under HMFOR-relevant conditions.

Overall, the results presented herein demonstrate that the SA-BPM integrated HER//HMFOR system provides a highly effective and scalable pathway to simultaneously achieve low-voltage hydrogen production and high-value chemical synthesis. This simple and robust coupling strategy holds great promise for enhancing both the energy conversion efficiency and economic competitiveness of next-generation electrochemical hydrogen production systems.

4 Conclusions

In this work, we developed a structurally self-anchoring bipolar membrane (SA-BPM) featuring an interlocked interfacial architecture to overcome the inherent limitations of conventional BPMs in water electrolysis. The introduction of a sub-micron anchoring morphology at the AEL–CEL junction significantly improved interfacial adhesion, increased active WD sites, and enhanced mechanical integrity under high current operation. The resulting SA-BPM demonstrated a low transmembrane resistance (4.8 Ω), a reduced onset voltage (0.79 V), and exceptional operational durability over 110 hours, even at high current densities (up to 1000 mA cm⁻²). Furthermore, coupling the SA-BPM system with anodic 5-HMF oxidation enabled efficient co-generation of hydrogen and 2,5-furandicarboxylic acid (FDCA), achieving high product selectivity and energy savings.

These findings underscore the importance of physical interface engineering in membrane functionality and provide a scalable strategy for designing next-generation BPMs tailored for multifunctional electrosynthesis platforms. The SA-BPM design paradigm offers new opportunities for integrating renewable hydrogen production with value-added chemical conversion in a single electrochemical system.

Author contributions

S. Li: data curation, formal analysis, investigation, writing; J. Jiang.: methodology, technical guidance; W. Wu.: resources; C. Wu.: conceptualization, funding acquisition, writing – review and editing; C. Yang.: project administration.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d5se01360g.

Acknowledgements

The authors acknowledge the financial support of this work from the National Natural Science Foundation of China (No. 22308363) and CNPC Innovation Fund (2024DQ02-0313).

References

1. M. El-Adawy, I. Dalha, M. Ismael, Z. Al-Absi and M. Nemitallah, Energy Fuel., 2024, 38, 22686–22718.
2. W. Suwaileh, Y. Bicer, S. Al Hail, S. Farooq, R. Mohamad Yunus, N. N. Rosman and I. Karajagi, Int. J. Hydrog. Energy, 2025, 121, 304–325.
3. A. D. Igalavithana, S. You, L. Zhang, J. Shang, J. Lehmann, X. Wang, Y.-G. Zhu, D. C. W. Tsang, Y.-K. Park, D. Hou and Y. S. Ok, ACS ES&T Eng., 2022, 2, 1987–2001.
4. W. Zhou, S. Li, Y. Yang, J. Yao, P. Chen, J. Liu, Y. Wu, Z. Li and F. Jin, Next Energy, 2025, 8, 100270.
5. M. Li, X. Deng, K. Xiang, Y. Liang, B. Zhao, J. Hao, J. L. Luo and X. Z. Fu, ChemSusChem, 2020, 13, 914–921.
6. J. Cheng, Y. Xiang, X. Huang and Z. Wei, Precis. Chem., 2024, 2, 447–470.
7. W. Li, N. Jiang, B. Hu, X. Liu, F. Song, G. Han, T. J. Jordan, T. B. Hanson, T. L. Liu and Y. Sun, Chem, 2018, 4, 637–649.
8. X. Gao, H. Yang, J. Qiu, L. Liu and J. Peng, ACS Appl. Mater. Interfaces, 2024, 16, 43582–43590.
9. J. Zeng, W. Chen, G. Zhang, L. Yu, L. Zhong, Y. Liu, S. Zhao and Y. Qiu, ACS Appl. Nano Mater., 2022, 5, 7321–7330.
10. P. K. Giesbrecht and M. S. Freund, Chem. Mater., 2020, 32, 8060–8090.
11. J. Yan, W. Yu, Z. Wang, L. Wu, Y. Wang and T. Xu, Aggregate, 2024, 5, e527.
12. M. A. Blommaert, D. Aili, R. A. Tufa, Q. Li, W. A. Smith and D. A. Vermaas, ACS Energy Lett., 2021, 6, 2539–2548.
13. Y. Xie, Z. Xu, Q. Lu and Y. Wang, Chin. J. Catal., 2024, 59, 82–96.
14. É. Lucas, J. C. Bui, T. N. Stovall, M. Hwang, K. Wang, E. R. Dunn, E. Spickermann, L. Shiau, A. Kusoglu, A. Z. Weber, A. T. Bell, S. Ardo, H. A. Atwater and C. Xiang, ACS Energy Lett., 2024, 9, 5596–5605.
15. K. Khoiruddin, I. G. Wenten and U. W. R. Siagian, Langmuir, 2024, 40, 9362–9384.
16. S. E. Renfrew, D. E. Starr and P. Strasser, ACS Catal., 2020, 10, 13058–13074.
17. F. Ye, S. Zhang, Q. Cheng, Y. Long, D. Liu, R. Paul, Y. Fang, Y. Su, L. Qu, L. Dai and C. Hu, Nat. Commun., 2023, 14, 2040.
18. G. Feng, Y. Pan, D. Su and D. Xia, Adv. Mater., 2024, 36, 2309715.
19. Z. Ma, L. Wang, G. Li and T. Song, Catalysts, 2024, 14 DOI:10.3390/catal14020157.
20. L. Chen and Y. Liu, Chem. Rec., 2025, 25, e202400238.
21. R. A. Tufa, M. A. Blommaert, D. Chanda, Q. Li, D. A. Vermaas and D. Aili, ACS Appl. Energy Mater., 2021, 4, 7419–7439.
22. A. Hohenadel, D. Powers, R. Wycisk, M. Adamski, P. Pintauro and S. Holdcroft, ACS Appl. Energy Mater., 2019, 2, 6817–6824.
23. E. Al-Dhubhani, H. Swart, Z. Borneman, K. Nijmeijer, M. Tedesco, J. W. Post and M. Saakes, ACS Appl. Energy Mater., 2021, 4, 3724–3736.
24. M. A. Blommaert, S. Subramanian, K. Yang, W. A. Smith and D. A. Vermaas, ACS Appl. Mater. Interfaces, 2022, 14, 557–563.
25. Y. Jeon, V. D. C. Tinh, V. D. Thuc and D. Kim, Chem. Eng. J., 2023, 459, 141467.
26. Z. Xu, L. Wan, Y. Liao, M. Pang, Q. Xu, P. Wang and B. Wang, Nat. Commun., 2023, 14, 1619.
27. Z. Xu, Y. Liao, M. Pang, L. Wan, Q. Xu, Y. Zhen and B. Wang, Energy Environ. Sci., 2023, 16, 3815–3824.
28. W. Zhu, H. Wang, G. Jing, Q. Liu, C. Li and H. Meng, RSC Adv., 2017, 7, 36313–36318.
29. E. Al-Dhubhani, M. Tedesco, W. M. de Vos and M. Saakes, ACS Appl. Mater. Interfaces, 2023, 15, 45745–45755.
30. S. Kole, G. Venugopalan, D. Bhattacharya, L. Zhang, J. Cheng, B. Pivovar and C. G. Arges, J. Mater. Chem. A, 2021, 9, 2223–2238.
31. M. A. Shehzad, A. Yasmin, X. Ge, Z. Ge, K. Zhang, X. Liang, J. Zhang, G. Li, X. Xiao, B. Jiang, L. Wu and T. Xu, Nat. Commun., 2021, 12, 9.
32. B. Eswaraswamy, P. Mandal, P. Goel and S. Chattopadhyay, ACS Appl. Polym. Mater., 2021, 3, 6218–6229.
33. N. Chen, C. Lu, Y. Li, C. Long, Z. Li and H. Zhu, J. Membr. Sci., 2019, 588, 117120.
34. C. Lu, C. Long, Y. Li, Z. Li and H. Zhu, J. Membr. Sci., 2020, 598, 117797.
35. K. Żebrowska, E. Coy, K. Synoradzki, S. Jurga, P. Torruella and R. Mrówczyński, J. Phys. Chem. B, 2020, 124, 9456–9463.
36. Y. Mao, W. Jiang, S. Xuan, Q. Fang, K. C.-F. Leung, B. S. Ong, S. Wang and X. Gong, Dalton Trans., 2015, 44, 9538–9544.
37. G. Ding, H. Lee, X. Cao, Z. Li, J. Liu, L. Wang, Y. Ding, L. Yin and L. Sun, ACS Energy Lett., 2025, 10, 571–578.
38. S. Xie, H. Fu, L. Chen, Y. Li and K. Shen, Sci. China Chem., 2023, 66, 2141–2152.
39. S. Favero, I. E. L. Stephens and M.-M. Titirci, Adv. Mater., 2024, 36, 2308238.
40. N. Chen, C. Lu, Y. Li, C. Long and H. Zhu, J. Membr. Sci., 2019, 572, 246–254.
41. E. Hong, Z. Yang, H. Zeng, L. Gao and C. Yang, ACS Mater. Lett., 2024, 6, 1623–1648.
42. Z. Ge, M. A. Shehzad, L. Ge, Y. Zhu, H. Wang, G. Li, J. Zhang, X. Ge, L. Wu and T. Xu, ACS Appl. Energy Mater., 2020, 3, 5765–5773.
43. Y. Chen, J. A. Wrubel, W. E. Klein, S. Kabir, W. A. Smith, K. C. Neyerlin and T. G. Deutsch, ACS Appl. Polym. Mater., 2020, 2, 4559–4569.
44. R. Pärnamäe, S. Mareev, V. Nikonenko, S. Melnikov, N. Sheldeshov, V. Zabolotskii, H. V. M. Hamelers and M. Tedesco, J. Membr. Sci., 2021, 617, 118538.
45. W. Yu, Z. Zhang, F. Luo, X. Li, F. Duan, Y. Xu, Z. Liu, X. Liang, Y. Wang, L. Wu and T. Xu, Nat. Commun., 2024, 15, 10220.
46. D. Chen, Y. Ding, X. Cao, L. Wang, H. Lee, G. Lin, W. Li, G. Ding and L. Sun, Angew. Chem., Int. Ed., 2023, 62, e202309478.
47. S. Cui, F. Wang, G. Wang, T.-T. Li, Z. Liu and Y. Wang, ACS Sustainable Chem. Eng., 2024, 12, 11767–11776.
48. L. Zeng, Y. Chen, M. Sun, Q. Huang, K. Sun, J. Ma, J. Li, H. Tan, M. Li, Y. Pan, Y. Liu, M. Luo, B. Huang and S. Guo, J. Am. Chem. Soc., 2023, 145, 17577–17587.
49. X. Chen, Y. He, G. Ding, Z. Feng, Q. Wei, C. Sui, Z. Wang and Q. Jiang, Nano Lett., 2025, 25, 9508–9515.
50. Z. Cheng, D. Fu, W. Zhou, W. Deng, Y. Tan and M. Ma, J. Mater. Chem., 2023, 11, 25983–25991.
51. D. Chen, Y. Ding, X. Cao, L. Wang, H. Ling, G. Lin, W. Li and G. Ding, Angew. Chem., Int. Ed., 2023, 62, e20230947.

---